A chloroplast cyclophilin functions in the assemblyand maintenance of photosystem II inArabidopsis thalianaAigen Fu, Zengyong He, Hye Sun Cho, Amparo Lima, Bob B. Buchanan*, and Sheng Luan*

Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720

Contributed by Bob B. Buchanan, August 20, 2007 (sent for review June 5, 2007)

Photosynthetic light reactions rely on the proper function of largeprotein complexes (including photosystems I and II) that reside inthe thylakoid membrane. Although their composition, structure,and function are known, the repertoire of assembly and mainte-nance factors is still being determined. Here we show that animmunophilin of the cyclophilin type, CYP38, plays a critical role inthe assembly and maintenance of photosystem II (PSII) supercom-plexes (SCs) in Arabidopsis. Mutant plants with the CYP38 geneinterrupted by T-DNA insertion showed stunted growth and werehypersensitive to high light. Leaf chlorophyll fluorescence analysisand thylakoid membrane composition indicated that cyp38 mutantplants had defects in PSII SCs. Sucrose supplementation enabledthe rescue of the mutant phenotype under low-light conditions,but failed to mitigate hypersensitivity to high-light stress. Proteinradiolabeling assays showed that, although individual thylakoidproteins were synthesized equally in mutant and wild type, theassembly of the PSII SC was impaired in the mutant. In addition, theD1 and D2 components of the mutant PSII had a short half-lifeunder high-light stress. The results provide evidence that CYP38 isnecessary for the assembly and stabilization of PSII.

thylakoid lumen � immunophilin � photosynthesis � protein folding �chaperone

The light reactions and attendant evolution of oxygen inphotosynthesis are carried out by four multisubunit protein

complexes residing in the chloroplast thylakoid membranes:photosystems I (PSI) and II (PSII), cytochrome b6f complex, andCFO–CF1 complex (1–3). For a complete understanding of thephotosynthetic process, it is essential to understand the biogen-esis and maintenance of the participating complexes. Earlierstudies on thylakoid protein supercomplex (SC) assembly, es-pecially PSII, concentrated on the role of stromal factors, suchas the translation and import machinery (4), because only alimited number of proteins were known to reside in the thylakoidlumen. However, recent proteomic findings suggest a populationof 80–100 proteins in that compartment (5–7). The immunophi-lin family is one of the predominant groups identified.

Immunophilins were originally discovered in their capacity ascellular receptors for immunosuppressive drugs: cyclosporin Aand FK506 (8, 9). The receptors for cyclosporin A and FK506,named cyclophilins (CYPs) and FK506-binding proteins (FK-BPs), respectively, were collectively designated as immunophil-ins. A common feature of most immunophilins is the associatedpeptidyl-prolyl cis-trans isomerase activity that catalyzes thecis-trans conversion of X-Pro peptide bonds, a rate-limiting stepin protein folding (8). These proteins are now known to occurwidely in organisms ranging from bacteria and fungi to animalsand plants. Studies in animal and plant systems have uncovereddiverse functions of immunophilins, such as protein foldases,chaperones, and scaffolding facilitators. They also possibly haveunknown catalytic capabilities (10, 11).

In Arabidopsis, genes encoding 52 putative immunophilinshave been identified: 23 FKBPs and 29 CYPs (12). This groupis by far the largest immunophilin family identified in any

organism. A striking feature of Arabidopsis immunophilin familyis that a large fraction is localized in the chloroplast, possiblyexplaining why the immunophilin family is more extensive inplants than animals. At least 16 immunophilins (11 FKBPs and5 CYPs) were identified in the thylakoid lumen (5, 6, 12), withlittle known of their function. The presence of such a largenumber of immunophilins with protein-folding capability sug-gests that these enzymes play a central role in the assembly andmaintenance of protein complexes, such as the two photosystemswhose subunits reside in or at least contact the thylakoid lumen.This possibility has been strengthened by a recent study showinga requirement for one of the FKBPs, 20-2, for the maintenanceof PSII (13). The present study shows that a member of the otherimmunophilin family in the thylakoid lumen, a CYP, is essentialfor assembling and stabilizing PSII. The results provide evidencefor a function of a CYP in chloroplasts.

Results and DiscussionIdentification of cyp38 Mutant Alleles. To identify their function, weisolated multiple Arabidopsis T-DNA insertion mutants for eachimmunophilin in the thylakoid lumen and screened them forpotential defects in growth under greenhouse conditions (long-day cycle with light intensity �150 �mol�m�2�s�1). Three T-DNAinsertion alleles for the gene encoding a 38-kDa CYP, CYP38(At3g01480), were obtained. One allele, referred to as cyp38-1,contained a T-DNA insertion in the promoter region of the gene.The second allele, cyp38-2, had a T-DNA insertion in the sixthexon of the coding region (�2,236 bp), and a T-DNA insertionwas found in the seventh exon (�2,506 bp) in the third allele(cyp38-3) (Fig. 1A).

RT-PCR analysis showed that the promoter insertion line(cyp38-1) had a significantly reduced level of the CYP38 tran-script (Fig. 1B). The mRNA of CYP38 was not detectable byusing RT-PCR in either cyp38-2 or cyp38-3, indicating that thesetwo alleles are null mutants. All three mutant alleles showedreduced growth and pale green leaves compared with wild typewhen grown under low-light conditions (25–35 �mol�m�2�s�1),and the phenotype of the cyp38-2 and cyp38-3 null mutants wasmore severe than that of the cyp38-1, which was concluded to bea weak mutant (Fig. 1C). The chlorophyll content in the nullmutants is �60% of the wild type based on fresh weight.

To demonstrate that the phenotype of cyp38-2 was attributableto the mutation of the CYP38 gene, we conducted complemen-tation experiments with the cyp38-2 mutant as background by

Author contributions: A.F. and Z.H. contributed equally to this work; A.F., Z.H., B.B.B., andS.L. designed research; A.F., Z.H., H.S.C., and A.L. performed research; A.F., Z.H., B.B.B., andS.L. analyzed data; and A.F., Z.H., B.B.B., and S.L. wrote the paper.

The authors declare no conflict of interest.

Abbreviations: PSI, photosystem I; PSII, photosystem II; SC, supercomplex; CYP, cyclophilin;FKBP, FK506-binding protein; NPQ, nonphotochemical quenching; BN, blue native.

*To whom correspondence may be addressed. E-mail: view@nature.berkeley.edu orsluan@nature.berkeley.edu.

© 2007 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0707851104 PNAS � October 2, 2007 � vol. 104 � no. 40 � 15947–15952

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

1

using a 6-kb genomic fragment containing a 2.5-kb promoterregion, the 2.6-kb full-length gene sequence, and the 900-bpdownstream sequence. The transgenic plants (cyp38-2C) dis-played wild-type growth, thus confirming that growth defects incyp38-2 resulted from mutation of the CYP38 gene (Fig. 1C). Wealso carried out a complementation test with a CYP38 cDNA andfound that it rescued the mutant as well (data not shown).

During the isolation of homozygotes from the cyp38-2 andcyp38-3 mutants, we noticed that the mutants were sensitive tohigh-light stress and died at an early stage unless cultivated underlow light. To investigate the light sensitivity of these mutants,wild-type and mutant plants were transferred to high light afterbeing grown initially for 7 weeks under low light. Two days afterexposure to 300 �mol�m�2�s�1 light, the mature leaves of the nullmutants (cyp38-2 and cyp38-3) started to wilt, whereas the weakmutant (cyp38-1) showed a typical high-light stress phenotype,shrinking and becoming purplish. Wild-type and complementedplants also showed a slightly purplish phenotype because of theaccumulation of anthocyanins (14), but they grew well nonethe-less (Fig. 1D).

Previous proteomic data suggested that CYP38 is included inthe chloroplast luminal proteome (5, 6). Its orthologue inspinach, TLP40, also was found to be localized in the chloroplastlumen and, further, was proposed to function in the phospory-lation of thylakoid proteins (15–17). The CYP38 also shareshomology with PAP-I, a protein purified from the pea and washypothesized to be an essential factor for RNA poly(A) poly-merase (18). Amino acid sequence analysis identified severalrelated proteins from cyanobacteria and algae, but not from

fungi or animals, suggesting conservation among photosyntheticorganisms. The protein features a C-terminal CYP domain anda leucine zipper at the N-terminal half (12, 15).

Because the two null alleles showed a similar phenotype,which was more severe than that of the weak allele, we chose oneof them, cyp38-2, for subsequent experiments to dissect thedefects in the cyp38 mutants. Results from experiments charac-terizing the mutant (phenotype, 2D gels, and immunoblots) wereconsistently similar with cyp38-3 (data not shown).

Chlorophyll Fluorescence Analysis Reveals Defects in PSII Function inthe cyp38 Mutants. The mutants’ stunted growth and sensitivity tohigh light suggested a possible defect in photosynthesis. Topursue this possibility, we measured chlorophyll f luorescence toassess photosynthetic electron transport in the leaves of wholeplants. The Fv/Fm ratio, which represents the maximum effi-ciency of PSII photochemistry, was remarkably reduced in themutant (0.65 in the mutant vs. 0.82 in the wild type), suggestingthat the PSII is photodamaged or misassembled even underlow-light growth conditions. �PSII, the quantum efficiency ofPSII photochemistry at different photon flux densities (19), alsowas decreased in the mutant (Fig. 2 Top), indicating a lowerefficiency of photosynthesis and thus suppressed growth. Non-photochemical quenching (NPQ), a reflection of the plant’sability to dissipate excess light energy as heat (20), was found tobe similar in mutant and wild-type plants under low-light inten-sities (�200 �mol�m�2�s�1) (Fig. 2 Middle). However, at lightintensities of �400 �mol�m�2�s�1, the NPQ of the mutant wassignificantly higher than that of the wild type. Further analysisrevealed that the additional NPQ was not rapidly reversible,

Fig. 1. Genetic characterization and phenotype of cyp38 mutant plants. (A)Localization of T-DNA insertion sites in CYP38 genomic DNA of three cyp38alleles: cyp38-1, cyp38-2, and cyp38-3. The CYP38 genomic DNA includes sevenexons (black boxes) and six introns (bold lines between black boxes). ATG,initiation codon; TAA, stop codon. (B) CYP38 mRNA levels in wild type (WT),cyp38-1, cyp38-2, and cyp38-3 as shown by RT-PCR. (C) Phenotype of wild type,cyp38-1, cyp38-2, cyp38-3, and cyp38-2 transformed with CYP38 genomic DNA(cyp38-2C) grown for 7 weeks under low light (25–35 �mol�m�2�s�1) with along-day cycle (16–8 h). (D) The plants shown in C were exposed to illumina-tion (300 �mol�m�2�s�1) for 2 days after 7 weeks of growth in low light (25–35�mol�m�2�s�1).

Fig. 2. Chlorophyll fluorescence analyses of wild-type (WT) and cyp38-2plants. Plants were grown under 25–35 �mol�m�2�s�1 for 7 weeks. Datarepresent means � SE (n � 6). �PSII, efficiency of PSII photochemistry; 1-qL,parameter estimating the fraction of PSII in close states or excitation pressureof PSII based on a lake mode (21).

15948 � www.pnas.org�cgi�doi�10.1073�pnas.0707851104 Fu et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

1

indicating that the increased NPQ is due to a larger componentof photoinhibitory quenching in the mutant. The energy-dependent quenching was similar in the mutant and wild type(data not shown). The parameter, 1-qL, which is considered tobe a better parameter than 1-qP to reflect the redox state of theQA electron acceptor of PSII (21), was lower in the mutant (Fig.2 Bottom). This result indicated a more oxidized plastoquinonepool in the mutant, a response likely due to a deficiency of PSIIrather than downstream defects. Taken together, the chlorophyllf luorescence analyses suggested that the cyp38 mutation affectedPSII function.

The PSII SC Is Compromised in cyp38 Mutants. The defects inphotosynthesis displayed by the cyp38 mutants were possiblycaused by a reduced level or malfunction in protein complexesin the electron transport chain. To examine this possibility, weanalyzed mutant and wild-type thylakoid membranes by using2D gel electrophoresis involving blue native (BN)/PAGE fol-lowed by SDS/PAGE.

BN gels developed with mutant and wild-type preparationscontaining equal amounts of chlorophyll revealed significantdifferences in the accumulation of PSII SCs, which were obviousin the wild type but almost absent in the mutant (Fig. 3A), thusconfirming defects in the PSII of mutant plants. By contrast, theLHCII monomer band was more abundant in the mutant, whichis not surprising because equal chlorophyll amounts were loadedonto the gel. Furthermore, the band corresponding to themonomeric PSI and CFO–CF1 complex was more abundant inthe mutant.

The SDS/PAGE gel following the BN gel revealed details inindividual protein components of the SC of the mutant and wildtype that were not apparent in the BN gel. For instance, themutant contained only trace amounts of PSII core proteins,whereas these components were abundant in the wild type (PSIISC in Fig. 3). In the right center area of the gel, mutant plantshad more of the components corresponding to LHCII trimer andunassembled proteins than the wild type. Thus, LHCII proteinswere distributed much less in the left, PSII SC (assembled) areaand more in the low molecular weight (unassembled) area,especially in the LHCII monomer form.

To analyze protein differences between the mutant and wildtype further, we applied immunoblotting to assess key compo-nents of each of the four major complexes of the chloroplastthylakoid membrane quantitatively: PSII, PSI, cytochrome b6f,and CFO–CF1. We prepared total protein and thylakoid mem-brane fractions from both mutant and wild type and probed themwith antibodies against subunits of each complex: D2 and D1 forPSII, CF1� for CFO–CF1, PsaD and PsaF for PSI, and Cytf andRieske (PetC) for Cytb6f complex.

The results summarized in Fig. 4 showed that the D2 and D1proteins of PSII were reduced in both the total protein (T) andmembrane (M) fractions of the mutant. However, although thelevel in the total protein fraction appeared not to change, the PSIsubunits of the mutant, PsaD and PsaF, were significantlyreduced in thylakoid membranes. The level of CF1� in themembrane fraction of the mutant also was decreased, althoughthe level of total protein appeared not to change. The Cytb6fcomplex displayed similar patterns at the level of Cytf and Rieskesubunits in the mutant and wild type. This result, which generally

Fig. 3. BN/PAGE and subsequent analysis of thylakoid membrane complexes in 2D SDS/PAGE. (A) Thylakoid membranes (5 �g of chlorophyll) were solubilized with1% dodecyl-�-maltoside and separated by 5–13.5%. Native gel slices were horizontally laid on the top of 2D SDS/PAGE. PSI-M, PSI monomer; PSII-D, PSII dimer; LHCII,PSII light-harvesting complex; LHCII-T, PSII light-harvesting complex trimer; LHCII monomer, PSII light-harvesting complex monomer. (B) Thylakoid membranecomplexes separated by BN/PAGE in A were further subjected to 12.5% SDS/PAGE and stained with silver. Protein identification was based on ref. 40.

Fig. 4. Immunodetection of thylakoid membrane proteins. Protein samples(1 �g of chlorophyll) were resolved by 12.5% SDS/PAGE, transferred to anitrocellulose membrane, and probed with antibodies against the indicatedthylakoid membrane proteins. WTT, total protein sample from wild-typeplants; 38-2T, total protein sample from cyp38-2; WTM, thylakoid membranesample from wild-type plants; 38-2M, thylakoid membrane sample fromcyp38-2.

Fu et al. PNAS � October 2, 2007 � vol. 104 � no. 40 � 15949

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

1

agrees with the 2D BN/SDS analysis, suggests that the PSII, PSI,and CF1 complexes in the mutant were compromised to variousdegrees, whereas the Cytb6 f complex was unaffected. Consistentwith the in vivo chlorophyll f luorescence analyses, the resultssuggest that the cyp38 mutants had primary defects in PSII,whereas the altered levels of PSI and CF1 subunits were likelydue to secondary effects of PSII deficiencies as described inother PSII mutants (22, 23).

cyp38 Mutants Display Defects in the Assembly of PSII SCs. Thereduced level of PSII SCs in the mutants could be attributed to(i) reduced synthesis or assembly or both, or (ii) increaseddegradation. We followed several approaches to distinguishthese possibilities.

Clearly, cyp38 mutants showed both severe defects in compo-nents of their thylakoid membranes and lower photosyntheticactivity. Thus, the stunted phenotype of the mutant as well as itshypersensitivity to high light could be due to secondary effectsof the mutation resulting from insufficient energy needed fornormal nourishment. To distinguish between primary defectsdue to lack of CYP38 and secondary defects caused by adeficiency in the products of photosynthesis, we supplied wild-type and mutant plants with 2% sucrose as a source of carbon tosupport normal growth and development.

Fig. 5A shows that, when grown on half-strength Murashige-Skoog (MS) medium for 4 weeks under low light without sucrose,cyp38-2 plants grew much less than their wild-type counterparts(Fig. 5A). By contrast, in the presence of 2% sucrose, both planttypes grew normally (Fig. 5B). However, when light intensity wasincreased, the mutant plants shown in Fig. 5B became bleached(Fig. 5C). Thus, while rescuing the stunted growth phenotype ofthe mutant, the addition of sucrose failed to mitigate its sensi-

tivity to high-light stress, suggesting that this defect is causedprimarily by a lack of the CYP38 protein.

To identify the PSII defects, we conducted BN/PAGE analysis ofthe thylakoid membranes from the cyp38 and the wild-type plantsgrown with 2% sucrose under low light. The gels revealed thatdifferences still existed (Fig. 5D) but to a much lesser degreecompared with soil-grown plants (Fig. 3A); that is, the mutantshowed a reduced level of PSII-LHCII SCs (PSII SC bands 1, 2, and5) and of the band containing PSI monomer � PSII dimer (PSI,PSII-D). Interestingly, immunoblot analysis revealed no significantdifferences between wild-type and cyp38-2 mutant plants in thecontent of components of the PSII, PSI, CFO-CF1, and Cytb6 fcomplexes: D1, PsaD � PsaF, CF1�, and Cytf � Rieske (Fig. 5E).This similarity suggested that the reduced level of certain individualsubunits observed in the soil-grown plants was caused by secondaryeffects resulting from reduced nutrition status. The analyses indi-cated that, provided a source of carbon, the cyp38 mutant resem-bled the wild type in being able to synthesize individual proteincomponents of the photosynthetic membrane complexes. However,unlike the wild type, the mutant did not assemble the PSII SCs inthe same pattern as the wild type, thereby leading to decreasedaccumulation of the complexes.

To distinguish between synthesis and assembly, we monitoredthe biosynthesis and assembly of subunit proteins in photosyn-thetic complexes directly by radiolabeling. Detached primaryleaves from sucrose-grown plants were fed [35S]Met, and cyclo-heximide was applied to inhibit the synthesis of nucleus-encodedproteins so that only chloroplast-encoded proteins were labeled(22). When the thylakoid membranes were resolved by SDS/PAGE, the protein patterns of the mutant and wild-type plantswere similar with respect to content and density of the individualbands (Psa A � B, CF1�, CF1�, CP47, CP43, D2, and D1) (Fig.6A). However, when the membrane preparations were analyzedby BN/PAGE, much lower levels of newly formed PSII SCs wereobserved in the cyp38 mutant relative to wild type, especially inthe gel exposed to film (Fig. 6B). In addition, a band corre-sponding to PSII dimer (PSII-D) in the wild type was essentiallymissing in the mutant. Based on the protein radiolabeling results,we concluded that the ability of the mutant to synthesize

Fig. 5. Effect of sucrose treatment on wild-type and cyp38-2 plants. (A)Wild-type and cyp38-2 plants were grown vertically on half-strength MSmedium without sucrose for 4 weeks under low-light conditions (25–35�mol�m�2�s�1). (B) Wild-type and cyp38-2 plants were grown vertically onhalf-strength MS medium supplemented with 2% sucrose under low-lightconditions (25–35 �mol�m�2�s�1) for 4 weeks. (C) Plants grown as described inB were exposed to high light (300 �mol�m�2�s�1) for 5 days. (D) BN/PAGE ofthylakoid membrane complexes from wild-type (WT) and cyp38-2 plants after4 weeks of growth on 2% sucrose half-width MS medium. PSI-M, PSI monomer;PSII-D, PSII dimer. (E) Immunoanalysis of thylakoid proteins (1 �g of chloro-phyll) from wild-type total protein (WTT), cyp38-2 total protein (38-2T),wild-type membrane protein (WTM), and cyp38-2 membrane protein (38-2M).Plants were grown on 2% sucrose medium for 4 weeks.

Fig. 6. In vivo synthesis of plastid-encoded membrane proteins. (A) Primaryleaves from wild-type (WT) and cyp38-2 (38-2) sucrose-grown plants were fed[35S]Met in the presence of cycloheximide. After a 20-min incubation, thyla-koid membrane proteins (5 �g of chlorophyll) were isolated and separated by12.5% SDS/PAGE. Proteins were then transferred to a nitrocellulose mem-brane and visualized autoradiographically. (B) After 20 min of labeling leaveswith [35S]Met (as in A), thylakoid membrane complexes (5 �g of chlorophyll)were separated by BN/PAGE. (Left) BN/PAGE pattern. (Right) Autoradiographyimage of the BN/PAGE. PSI-M, PSI monomer; PSII-D, PSII dimer; RC47, PSIIreaction center complex without CP43 subunit (41); RC, PSII reaction centercomplex without CP47 and CP43 subunits (41).

15950 � www.pnas.org�cgi�doi�10.1073�pnas.0707851104 Fu et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

1

thylakoid membrane proteins was similar to the wild type.However, the cyp38 mutant plants were impaired in the assemblyof the PSII SC.

cyp38 Mutants Also Display Defects in the Stability of PSII SCs. Thein vivo radiolabeling experiments demonstrated defects in thePSII assembly in the mutant, but did not exclude the possibilitythat the mutant PSII SCs had a shorter half-life compared withwild type. To test whether the CYP38 protein is important forPSII maintenance, we monitored the levels of several thylakoidmembrane proteins by immunoblotting. We used leaves fromsucrose-grown plants, in which protein biosynthesis was inhibitedwith lincomycin (chloroplast-encoded proteins) and cyclohexi-mide (nuclear-encoded proteins). Under these conditions, areduced level of a particular protein reflects its degradation.Under low light, the level of proteins examined (CF1�, D1,PsaD, PsaF, and Cytf ) did not significantly decrease in either themutant or wild type (Fig. 7A). By contrast, under high light (300�mol�m�2�s �1), the levels of the D1 and D2 components of PSIIwere dramatically reduced in the mutant, whereas the levels ofmembers of the other complexes were similar (monitored byCF1�, PsaD, PsaF, Cytf, and Rieske) (Fig. 7B). The resultssuggest that D1 and D2 were degraded to a greater extent in highlight in the absence of CYP38 possibly because of a defect in theassembly of the PSII SC. Therefore, they emerge as targetshighly susceptible to photodamage as in earlier studies (24, 25).

Concluding Remarks. Because the structures of the photosyntheticmembrane complexes of chloroplasts have been described inx-ray crystallography analyses (26, 27), it becomes increasinglyimportant to understand the mechanisms underlying their as-

sembly and maintenance. To this end, it is essential to identifyand understand the proteins that function in the assembly andmaintenance of the supermolecular complexes that catalyze thisenergy conversion in nature. Earlier work has shown thatproteins functional in the translation and import of subunitproteins are required for the biogenesis of photosynthetic com-plexes (4, 28). However, the proteins that function specifically inthe assembly of the complexes have long been a mystery and areonly now beginning to be identified (29–31).

Recent studies on immunophilins of the thylakoid lumen havebegun to uncover information central to the problem. Thus,potential functions of luminal immunophilins have been linkedto the biogenesis of photosynthetic complexes. One, FKBP13,interacts with the Rieske protein in the accumulation of thecytochrome b6f complex (32). Another, FKBP20-2, is requiredfor the accumulation of PSII SCs (13). Finally, as seen here, theother type of immunophilin, CYP38, is linked to the assemblyand maintenance of PSII. Owing to their extensive presence inthe thylakoid lumen, it seems likely that other immunophilinswill be found to participate in the biogenesis and maintenance ofphotosynthetic complexes.

Materials and MethodsPlant Materials and Growth Conditions. Wild-type (ecotype Colum-bia-0) and cyp38 T-DNA insertion Arabidopsis thaliana mutant(ecotype Columbia-0) seeds were obtained from Syngenta (Re-search Triangle Park, NC). For soil-grown plants, sown seedswere cold-treated for 2 days and then grown under long-dayconditions (16-h illumination of 25–35 �mol�m�2�s�1 and 8-hdark cycle). For plate-grown plants, seeds were surface-sterilizedwith 70% ethanol for 5 min and 50% bleach for 10 min and thenwashed four times with sterilized distilled water. The washedseeds were plated on 0.8% agar plates containing half-strengthMS medium with or without 2% sucrose under the same growthconditions as soil-grown plants.

Identification of T-DNA Insertional Mutants and Complementationwith Transgenic Plants. Three mutant lines were obtained fromSyngenta: Garlic 204 32, Garlic 206 38, and Garlic 321 44.Homozygous lines were isolated by confirming the T-DNA insertby PCR. DNA isolation and T-DNA border information wereobtained according to ref. 33. RT-PCR was used to verifyexpression levels in homozygous plants. RNA isolation, cDNAsynthesis, and RT-PCR were performed according to standardprocedures (34). For complementation by genomic DNA, a 6-kbgenomic DNA fragment containing the AtCYP38 gene and the2.5-kb upstream and 900-bp downstream sequences were ampli-fied by PCR and cloned behind the CaMV 35S promoter in thebinary vector pCAMBIA1300. For complementation by cDNA,the coding region of the gene was amplified by RT-PCR fromtotal RNA and cloned into pCAMBIA1301 without the GUSgene. These constructs were further transferred into cyp38-2plants by using the floral dip method (35).

Chlorophyll Fluorescence Measurements. Chlorophyll f luorescenceparameters were measured with an FMS2 fluorometer (Han-satech, Norfolk, UK) with intact leaves of wild-type and mutantplants grown for 7 weeks in soil under low-light conditions(25–35 �mol�m�2�s�1) (13). Fv/Fm was defined as (Fm �Fo)/Fm, and �PSII was defined as (F�m � Fs)/F�m. NPQ wascalculated as (Fm � F�m)/F�m; 1-qL is equal to (F�o/Fs)(F�m �Fs)/(F�m � F�o). Fm is the maximum fluorescence in the dark-adapted state; F�m is the maximum fluorescence in any light-adapted state; Fs is the steady-state fluorescence in the light; Fois the minimal fluorescence in the dark-adapted state; and F�o isthe minimal fluorescence in any light-adapted state.

Fig. 7. Half-life of thylakoid membrane proteins in wild-type and cyp38plants. (A) Arabidopsis leaves from sucrose-grown wild-type (WT) and cyp38-2plants were incubated with lincomycin and cycloheximide for 30 min andilluminated under 25–35 �mol�m�2�s�1 for 0, 2, and 4 h. Thylakoid membraneproteins (1 �g of chlorophyll) were resolved by SDS/PAGE, and the gel wasimmunoblotted to assess protein half-life. (B) Conditions as in A, except thatleaves were illuminated under 300 �mol�m�2�s�1 for 0 and 2 h.

Fu et al. PNAS � October 2, 2007 � vol. 104 � no. 40 � 15951

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

1

Preparation of Total Protein Sample and Thylakoid Membranes.Arabidopsis leaf tissues were homogenized at 4°C with a mortal andpestle in 0.33 M sorbitol/10 mM EDTA/50 mM Hepes (pH 8.0). Thehomogenate was filtered through two layers of Miracloth (Calbio-chem, La Jolla, CA), and a small portion of the filtrate was takenas a total protein sample. The remaining portion of the filtrate wasused to prepare thylakoid membranes (36).

Immunoblotting Analysis. Western blotting was performed asbefore (37). In brief, protein samples corresponding to equalamounts of chlorophyll were electrophoresed through 12.5%SDS/PAGE, and protein was transferred to a nitrocellulosefilter. The filter was incubated with antibodies generated againstthe indicated thylakoid membrane proteins and visualized byusing the ECL immunodetection procedure (Amersham Bio-sciences, Pittsburgh, PA).

Half-Life of Thylakoid Membrane Proteins. Expanding leaves from 4-to 5-week-old Arabidopsis plants (wild type and cyp38-2) grownon the MS plates with 2% sucrose were vacuum-infiltrated with100 �g/ml lincomycin and 20 �g/ml cycloheximide as indicatedto block chloroplast- and nucleus-encoded protein synthesis,respectively (38). After 30-min incubation, the leaves wereilluminated under growth conditions for 0, 2, and 4 h or under300 �mol�m�2�s�1 for 2 h. After treatment, thylakoid membranes

were isolated, and the content of the PSII proteins was deter-mined by immunoblot analysis.

BN/PAGE and 2D SDS/PAGE. BN/PAGE was performed accordingto refs. 13 and 39. For 2D analysis, excised BN/PAGE lanes weresoaked for 30 min in 2 SDS sample buffer supplied with 5%�-mercaptoethanol. Each lane with denatured proteins wasplaced on top of 1.5-mm-thick 12% SDS/PAGE and electro-phoresized at room temperature with a constant current of 20mA. After electrophoresis, gels were stained with silver (40).

In Vivo Radiolabeling of Chloroplast Proteins. Primary leaves fromplants supplied with sucrose for 4 weeks were used for in vivoprotein radiolabeling according to refs. 22 and 38, with minormodifications. Isolated thylakoid membranes were subjected toeither SDS/PAGE or BN/PAGE with equal chlorophyll loading.The SDS/PAGE was transferred to a nitrocellulose membraneand visualized autoradiographically. BN/PAGE were dried di-rectly with a gel drier (BioRad, Hercules, CA) and exposed tox-ray film.

We thank Drs. A. Melis, S. Park, and R. Malkin (University of California,Berkeley) for providing antibodies and K. Niyogi (University of Cali-fornia, Berkeley) for obtaining and interpreting the chlorophyll f luo-rescence data. This work was supported by U.S. Department of Agri-culture grants (to B.B.B. and S.L.) and U.S. Department of Energygrants (to S.L.).

1. Nelson N, Yocum CF (2006) Annu Rev Plant Biol 57:521–565.2. Barber J (2006) Biochem Soc Trans 34:619–631.3. Shi L, Schroder WP (2004) Biochim Biophys Acta 1608:75–96.4. Wollman FA, Minai L, Nechushtai R (1999) Biochim Biophys Acta 1411:21–85.5. Peltier JB, Emanuelsson O, Kalume DE, Ytterberg J, Friso G, Rudella A,

Liberles DA, Soderberg L, Roepstroff P, von Heijne G, von Wijk KJ (2002)Plant Cell 14:211–236.

6. Schubert M, Petersson UA, Haas BJ, Funk C, Schroder WP, Kieselbach T(2002) J Biol Chem 277:8354–8365.

7. Spetea C, Hundal T, Lundin B, Heddad M, Adamska I, Andersson B (2004)Proc Natl Acad Sci USA 101:1409–1414.

8. Schreiber SL (1991) Science 251:283–287.9. Luan S (1998) Bot Bull Acad Sin 39:217–223.

10. Buchanan BB, Luan S (2005) J Exp Bot 56:1439–1447.11. Romano P, Gray J, Horton P, Luan S (2005) New Phytol 166:753–769.12. He Z, Li L, Luan S (2004) Plant Physiol 134:1248–1267.13. Lima A, Lima S, Wong JH, Phillips RS, Buchanan BB, Luan S (2006) Proc Natl

Acad Sci USA 103:12631–12636.14. Winkel-Shirley B (2002) Curr Opin Plant Biol 5:218–223.15. Fulgosi H, Vener AV, Altschmied L, Herrmann RG Andersson B (1998)

EMBO J 17:1577–1587.16. Vener AV, Rokka A, Fulgosi H, Andersson B, Herrmann RG (1999) Bio-

chemistry 38:14955–14965.17. Rokka A, Aro EM, Herrmann RG, Andersson B, Vener AV (2000) Plant

Physiol 123:1525–1536.18. Gupta JD, Li Q, Thomson AB, Hunt AG (1998) Plant Mol Biol 37:729–734.19. Maxwell K, Johnson GN (2000) J Exp Bot 51:659–668.20. Muller P, Li XP, Niyogi KK (2001) Plant Physiol 125:1558–1566.21. Kramer DM, Johnson G, Kiirats O, Edwards GE (2004) Photosynth Res

79:209–218.

22. Meurer J, Plucken H, Kowallik K, Westhoff P (1998) EMBO J 17:5286–5297.23. Plucken H, Muller B, Grohmann D, Westhoff P, Eichaker LA (2002) FEBS Lett

532:85–90.24. Baena-Gonzalez E, Aro E (2002) Philos Trans R Soc London Ser B 357:1451–

1460.25. Melis A (1999) Trends Plants Sci 4:130–135.26. Nelson N, Ben-Sham A (2004) Nat Rev Mol Cell Biol 5:971–982.27. Merchant S, Sawaya MR (2005) Plant Cell 17:648–663.28. Jarvis P, Robinson C (2004) Curr Biol 14:1064–1067.29. Nakamoto SS, Hamel PP, Merchant S (2000) Biochimie 82:603–614.30. Rochaix JD (2002) FEBS Lett 529:34–38.31. Melkozernov AN, Barber J, Blankenship RE (2006) Biochemistry 45:331–345.32. Gupta R, Mould R, He Z, Luan S (2002) Proc Natl Acad Sci USA 99:15806–

15811.33. Sessions A, Burke E, Presting G, Aux G, McElver J, Patton D, Dietrich B, Ho

P, Bacwaden J, Ko C, et al. (2002) Plant Cell 14:2985–2994.34. Sambrook J, Russell DW (2001) Molecular Cloning: A Laboratory Manual (Cold

Spring Harbor Lab Press, Cold Spring Harbor, NY).35. Clough SJ, Bent AF (1998) Plant J 16:735–743.36. Chen M, Choi YD, Voytas D, Rodermel S (2000) Plant J 2:303–313.37. Fu A, Park S, Rodermel S (2005) J Biol Chem 280:42489–42496.38. Peng L, Ma J, Chi W, Guo J, Zhu S, Lu Q, Lu C, Zhang L (2006) Plant Cell

18:955–969.39. Asakura Y, Hirohashi T, Kikuchi S, Belcher S, Osborne E, Yano S, Terashima

I, Barkan A, Nakai M (2003) Plant Cell 16:201–214.40. Schwenkert S, Umate P, Dal Bosco C, Volz S, Mlcochova L, Zoryan M,

Eichacker LA, Ohad I, Herrmann RG, Meurer J (2006) J Biol Chem 281:34227–34238.

41. Rokka A, Suorsa M, Saleem A, Battchikova N, Aro EM (2005) Biochem J388:159–168.

15952 � www.pnas.org�cgi�doi�10.1073�pnas.0707851104 Fu et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

1